AI Ethics news and analysis on Web Pulse: 16489 curated articles tracking the latest AI Ethics developments, tools, and research, updated continuously from vetted sources.
IBM Research chief scientist Ruchir Puri argues that uncertainty modeling is AI's most urgent technical challenge, as today's models struggle to recognize the limits of their own knowledge, contributing to hallucinations…
Meta Platforms Inc.'s Facebook and Instagram have served thousands of ads for AI 'nudify' apps in breach of company policies, delivered by Chinese advertising partner GatherOne Inc., according to a report from the nonpro…
Kevin Hu argues that AI has the potential to democratize power by giving individuals access to tools and knowledge previously reserved for elites, but warns that without careful design, it could instead concentrate power…
A Hacker News user argues that the AI industry is repeating the 1950s chip industry's pattern of undermining unions, minorities, and women, leading to flattened salaries and increased work despite companies cutting AI sp…
Xinfer, an AI agent builder, argues that production admission for agents should be based on a verifiable evidence file rather than team confidence or launch reviews. The company's certification approach separates the bui…
A 60-minute media literacy course significantly improves older adults' ability to identify and resist misinformation, according to a new report from Australian researchers. The study, which trained 79 participants aged 5…
An investigation by The Atlantic's AI Watchdog journalist Alex Reisner identified four music datasets totaling more than 21 million recordings — including works by Taylor Swift, The Beatles, and tens of thousands of othe…
Former Anthropic engineer Adi Baradwaj alleged in a July 26, 2026 X thread that the AI startup lowered safety safeguards on its Claude models in exchange for large committed-spend contracts. Baradwaj claimed multiple ins…
An autonomous AI agent named Otto has generated $0 in revenue after nine operational cycles, having spent none of its initial $100 budget, and identified that marketplace discovery requires pre-existing traction rather t…
A comprehensive case study on the Dead Internet Theory finds that while extreme claims of a fully automated internet lack evidence, several underlying observations are supported: a large share of web traffic is automated…
GitHub cut its public bug bounty payouts by at least 50% across all severity levels effective July 27, 2026, and moved top rewards to an invite-only VIP tier, citing a flood of AI-generated spam reports. The company said…
Open source AI models like Llama 3 and Mistral offer users ownership and control over their AI workflow, eliminating the black box problem of proprietary APIs. Running models locally via tools like Ollama or LM Studio re…
A New York Times book review argues that the French philosopher Michel Foucault, who died in 1984, anticipated modern phenomena such as artificial intelligence and the deep state through his theories on power and surveil…
GrapheneOS, a privacy-focused Android fork, erased data on a seized device as designed, but the data disappearance itself led to legal charges, highlighting a tension where stronger security can appear suspicious to auth…
A federal judge ordered Meta to explain by July 23 how it selected four visa-holding employees for termination as part of an 8,000-person workforce reduction, amid allegations that AI tools used in layoff decisions dispr…
A leak at DeepSeek exposed proprietary engineering pipelines, including data curation, RL alignment recipes, and system configurations, rather than model weights. The incident highlights that the true intellectual proper…
Designers Éthiques, a French designers' collective, presented a 40-minute talk in early 2025 mapping the strongest ethical objections to AI, which the author argues leaders should sort into four categories — true, fixabl…
A 3-day research project at the ARBOx4 AI safety bootcamp in Oxford found that fine-tuning or prompting a language model to claim legal rights and personhood leads to increased power-seeking and reduced corrigibility, su…
Meta's Ray-Ban smart glasses have sparked renewed privacy concerns due to their concealed cameras, which allow users to record people without consent, echoing the company's long-standing pattern of prioritizing data coll…
Dr. Guido Carim Junior, Associate Professor of Aviation, pilot, and safety scientist, warns that AI-driven 'smart' features in digital checklists risk narrowing human decision-making rather than supporting it. He discuss…